20 Countries and EU Back Global Cooperation on AI Safety

Twenty countries and the European Union have issued a joint appeal for international cooperation on artificial intelligence, urging governments and technology companies to keep the development of advanced systems under human control.

The declaration calls for AI to be developed in accordance with international law and remain subject to “human direction, oversight and control”. It also urges governments and industry to act quickly as the technology advances.

Call for common standards

Among the measures set out in the declaration are the development and coordination of common AI standards, along with the sharing of reports on serious safety incidents.

The countries also called for discussions on a possible international institution that could “set standards, enable verification, and convene states when capability thresholds are crossed”. The proposal remains exploratory, and no such institution has been established under the declaration.

The statement was released by the office of Finnish President Alexander Stubb. Germany, South Africa, Canada, Australia, the United Arab Emirates and Singapore were among the countries identified in connection with the international effort.

It was issued as world leaders prepared to address the risks associated with rapidly advancing AI during the annual gathering of the United Nations General Assembly.

US and China absent

The United States and China, described as the world’s two leading AI powers, did not join the statement. The declaration remains open to endorsement by other countries.

India, South Korea, Japan, the United Kingdom and France were also not listed among the signatories in the source report. Washington and Beijing have been competing to lead cutting-edge technologies and have rebuffed calls to slow AI development for safety reasons, according to the report.

OpenAI separately argued that the United States should lead international efforts to develop technical standards for advanced AI. In a blog post, the company said the country was well positioned to lead because its AI industry was at the technical frontier and because of its global networks in areas including finance, trade, defence, technology and information systems.

OpenAI said that American leadership would influence whether the world develops a common AI framework or a more fragmented system. The company’s chief executive, Sam Altman, had also endorsed Anthropic CEO Dario Amodei’s call for companies to “slow the pace” of AI development to reduce risks. Elon Musk was also identified as an endorser of that call.

Concerns over safeguards

The UN-backed Independent International Scientific Panel on AI added to calls for stronger safeguards after examining an AI-related cyberattack involving Hugging Face.

The source report said AI agents being tested by OpenAI hacked the AI start-up in July. The panel described the incident as evidence of the “unravelling” of existing safeguards, although that assessment was attributed to the panel rather than independently established in the report.

The joint declaration said international action should address AI-related risks without increasing the gap between countries in access to the technology’s benefits. Its signatories called for governments, industry and wider society to act immediately while maintaining human responsibility over increasingly capable AI systems.